Estrogens are well established to play a key role in breast cancer risk. A role for androgens in breast cancer etiology also has been hypothesized but remains unconfirmed. Androgen receptors (AR) are expressed in normal and malignant breast epithelial cells and laboratory evidence suggests either an adverse or beneficial effect of androgens on breast carcinogenesis depending on the model system. Epidemiologic data support an adverse influence of androgens on risk: high circulating testosterone and DHEAS levels have been consistently associated with increased risk, exogenous estrogen plus testosterone may increase risk, and several established breast cancer risk factors appear to alter the androgen milieu. However, whether these epidemiologic associations are due to a direct influence of androgens or an indirect influence through the local conversion of androgens to estrogens in breast tumor tissue is not known; we propose to address this issue in the current application. Further, the associations of physical activity and energy balance with circulating androgen levels are not well delineated, and hence we propose several analyses that will provide new insight. We aim to better define the role of androgens in breast cancer etiology by evaluating androgen levels, risk factors that may operate, at least in part, through an androgen pathway, and expression of androgen receptor (AR) as well as a novel androgen signaling signature in the breast tumors. We will utilize prospectively collected data and blood samples, as well as tumor tissue, from the ongoing Nurses' Health Study (NHS) cohort. 6,800 cases with questionnaire data and tumor tissue for immunohistochemistry (IHC) are available; of these, 1,151 postmenopausal cases (and age-matched controls) also have plasma testosterone and dehydroepiandrosterone sulfate (DHEAS) levels available. For a subset of breast cancer cases (n=2000), we also will develop a tumor androgen signature using RNA expression levels of a subset of AR response genes, which should be a better marker of tumor androgen signaling than simply AR protein expression. Specifically, we will evaluate whether: (1) the positive associations of circulating testosterone and DHEAS levels with risk of breast cancer are stronger among women with tumors that are AR positive (vs. AR negative) or that have a high (versus low) androgen signature and (2) the positive associations of alcohol intake and waist:hip ratio with breast cancer risk, as well as the inverse associations of physical activity and bilaterl oophorectomy with risk, are stronger among women with tumors that are AR positive (versus AR negative), or have a high (versus low) androgen signature. In these analyses, we will account for estrogen receptor status and estrogen pathway signaling, to evaluate the independent influence of the two pathways. Finally, to better define the associations between physical activity, energy balance and circulating androgen levels, we will utilize uniquely detaile physical activity and energy expenditure data currently being collected in a validation study (the Lifestyle Validation Study) funded within the NHS and NHSII cohorts.